Pervasive Developmental Disorders (PDD) are serious developmental conditions affecting 6 per 1000 in school-age children. We propose a multisite study in which 120 children with PDD accompanied by tantrums, aggression, and/or self-injury will be randomly assigned to risperidone only (N=40) or risperidone plus a structured behavior therapy program (N=80) and then compared on adaptive and behavioral outcomes after 16 weeks. After an additional eight weeks of stabilization (6 months of treatment total), children who demonstrate a positive response to the combination of risperidone and behavior therapy (estimated samples size of 60) will be randomized to continued treatment with risperidone or gradual discontinuation. The goal of the discontinuation phase is to examine whether the addition of behavior therapy permits discontinuation of risperidone. The justification for the inclusion of children and adolescents with PDD is based on the observation that PDD can be a serious disability across social, emotional, and academic domains. Thus, there is a pressing need for new treatments in PDD that are safe and effective. In a previous RUPP Autism Network study, we showed that risperidone was safe and effective for children and adolescents with autism accompanied by tantrums, aggression, and/or self-injury. We also showed that the behavioral symptoms rapidly returned in a high percentage of these pediatric patients when the medication was withdrawn in a placebo-controlled discontinuation study. The present study will extend these findings and will test whether the combination of risperidone and a structured behavior therapy program will enhance adaptive functioning in children with PDD and protect against relapse. Side effects and therapeutic response will be carefully monitored. Subjects will be genotyped in order to evaluate the relationship of specific serotonin receptor genotypes and treatment response. The proposed study will be conducted by a multisite network that has already shown the capacity to collaborate in the design and conduct of clinical trials in the PDD population. Furthermore, this five-site network has established reliability on several measures that are relevant to the conduct of clinical trials in children with PDD.